diff --git a/examples/fairseq/models/language_modeling.py b/examples/fairseq/models/language_modeling.py
index 38cf9e2b..206de181 100644
--- a/examples/fairseq/models/language_modeling.py
+++ b/examples/fairseq/models/language_modeling.py
@@ -25,7 +25,7 @@
 
 from torchscale.architecture.config import DecoderConfig
 from torchscale.architecture.decoder import Decoder
-from torchscale.model.LongNet import LongNetDecoder
+from torchscale.model.longnet import LongNetDecoder
 
 DEFAULT_MAX_TARGET_POSITIONS = 1024
 logger = logging.getLogger(__name__)
diff --git a/examples/longvit/longvit.py b/examples/longvit/longvit.py
index 58db485b..322e250e 100644
--- a/examples/longvit/longvit.py
+++ b/examples/longvit/longvit.py
@@ -23,7 +23,7 @@
 import torch.nn as nn
 
 from torchscale.architecture.encoder import Encoder
-from torchscale.model.LongNet import LongNetEncoder
+from torchscale.model.longnet import LongNetEncoder
 from torchscale.architecture.config import EncoderConfig
 from timm.models.layers import trunc_normal_ as __call_trunc_normal_
 
diff --git a/examples/longvit/pretraining/vision_transformer.py b/examples/longvit/pretraining/vision_transformer.py
index 675ee8a9..0eb0cf0d 100644
--- a/examples/longvit/pretraining/vision_transformer.py
+++ b/examples/longvit/pretraining/vision_transformer.py
@@ -23,7 +23,7 @@
 
 from utils import trunc_normal_
 from torchscale.architecture.encoder import Encoder
-from torchscale.model.LongNet import LongNetEncoder
+from torchscale.model.longnet import LongNetEncoder
 from torchscale.architecture.config import EncoderConfig
 
 
diff --git a/torchscale/model/LongNet.py b/torchscale/model/longnet.py
similarity index 100%
rename from torchscale/model/LongNet.py
rename to torchscale/model/longnet.py